Fault prognosis for complex system based on Bayesian network

Abstract:
In order to study the fault prognosis method of complex systems, a Bayesian network model is established. The time information is incorporated into the network through a qualitative trend analysis method. The fault probability of a node is determined by combining the development trend and healthiness of node through the fuzzy affiliation function. The fault prediction results of complex systems under the complete data set and non-complete data and are obtained through parameter learning and multi-tree propagation algorithm, which verifies the accuracy and feasibility of the method for complex systems.